For us, two of the plagues hitting us were of our doing or under our control: the ZeBu-Server pre-announcement and the sales team over-growth. The economy was, of course, completely out of our hands. So there we were, stalled out at the top of our climb, and we were pulling every lever and pushing every button we could to right things and make sure that we stayed aloft, understanding that we were also flying in turbulent air, and there was nothing we could do about that.
EDA WEEKLY: So you focused only on what you could change.
Exactly. We solved what we could. ZeBu-Server was finally finished and made available for purchase, and the sales flow picked up again. We went back to a smaller sales team, only this time we focused hard on making sure they had the training and resources they needed to be successful. We will undoubtedly need to grow this sales team, but now we know how to do it.
Meanwhile the economy has been inching back, and the world is gaining some renewed confidence that, even if some of the past practices weren’t a good idea, at least the entire system hasn’t been scuttled. This reinvigorates investment, and, as a result, 2010 gave us the 50% year-to-year sales boost after a sobering 2009. While perhaps more modest than 100%, it’s a validation that our fundamentals are sound, and that we are now on a firmer footing than we were before the crisis.
EDA WEEKLY: So, for now, the turbulence is at least manageable?
From this point, our roadmap is clear. We have been successful riding the Virtex wave, and, as before, the upcoming rounds of Virtex will enable yet higher capacity and speed. We have a real opportunity to take ownership of a significant piece of the emulation pie.
EDA WEEKLY: But at the moment, the press is full of news about legal issues.
Yes. Success comes with its own challenges. You can fly under the radar with impunity, but once you get high enough, you attract attention. And not all attention is good. Large companies have lots of resources, both for technology development and for legal teams. Twice now we’ve run into legal challenges. The first was back in 2006; we made a tough decision to settle, even though we felt we could have won on the merits of our case. But that would have taken a lot of money and would have distracted from running the business, so we let discretion be the better part of valor and settled.
Today, as if hearing the echoes of the early parting of ways, we again see ourselves being challenged – this time for using the Xilinx readback feature that any designer could have used any time for many years. The first attempt to have imports to Japan shut down failed; now it’s moved from the Customs House to the courts. We believe strongly in our position, and we will need to defend ourselves in a manner that doesn’t take away from daily operations.
EDA WEEKLY: Let’s put the legal challenges aside for the moment. What other issues are on EVE’s plate now?
We see nothing but opportunity as technology moves to and beyond the 28-nm node. With so many gates on a chip, simulation becomes impossible as a full functional verification strategy.
Designers will, of course, still simulate to validate their RTL at the block level, but simulation will simply not be able to keep up with the integration of those blocks into complex SoCs.
Formal verification provided by EDA players like Real Intent, Jasper, and Atrenta, will continue their market penetration, albeit not as replacement for functional verification, but as a complementary verification methodology.
Beyond simple hardware verification, a larger and larger share of chip functionality is being carried out by software. And that software needs to be verified. Not just to prove that the software itself is correct, but also to ensure that the software executes properly on the underlying hardware, and that performance is as desired.
If high-level architectural tuning is required – multicore balancing, cache sizes, context swapping churn, bus contention – all critical real-world parameters that are hard to simulate, it will be more and more important that real software be executed on a real implementation of the design while it can still be changed. Discovering a mismatch between software requirements and hardware platform after silicon is in hand is not an option. Emulation – or co-emulation, along with a virtual model on the host – is the only practical way to validate software/hardware interaction prior to silicon.
All of this means that we have a lot of work to do for as far ahead as we can see.
EDA WEEKLY: So what is your prognosis now for EVE, assuming the legal issues facing the Company go away? How does the flight look from here on out?
Well, here we are, at 20,000 feet and climbing. We are far from reaching our cruising altitude, and it’s not really a good idea to unfasten our seatbelts and order that complementary cognac. On the other hand, we’ve survived some significant clear-air turbulence, we’ve tightened up our operation, and visibility is good.
We anticipate an excellent flight for our team, our partners, and our customers.
EDA WEEKLY: Good for you; we wish you success from now on. And congratulations in sticking with the airplane analogy; I have come to really like it!
 Footnote: The first EDA Weekly in the current series was devoted to, “The Role of Business Planning,” appearing initially on November 9, 2009.
Loyal readers of the writer’s earliest efforts to produce EDA WEEKLIES in this series, will recall that when it came time about a year ago to do an article on Altium Limited, the writer still harbored hopes of a trip to Altium corporate headquarters, i.e. to Australia:
In fact, Altium's world headquarters are located on the 'upper north shore' in Belrose, Australia, only 17 miles and 40 minutes by car from the Sydney city center.
In as much as trips by this writer to each company HQ's were part of the necessary activity to score most of the interviews associated with his EDA Weeklies up to that time, (e.g. hot spots like San Jose, Fremont, and Santa Rosa, CA), fantasizing about a sojourn to Sydney certainly seemed in order. Until of course it was mentioned that Sydney was 7,420 air miles from San Francisco (or 14,166 miles through Hawaii and Japan to Australia by sea kayak).